Description
This regular expression is designed to tokenize JSON-like content embedded in INI files. It is not a full JSON validator; instead, it provides a lightweight lexical scanner that identifies structural tokens (objects, arrays, strings, numbers, booleans, null) and ignores comments and whitespace. The token stream is then processed by a hand‑written recursive‑descent parser (with depth‑limit protection) to build a .NET object graph (Dictionary, object], or primitives).
Regex Pattern
(?//.|/\.?\/)|
(?""(?:\\.)"")(?=(?:\s|//.|/\.?\/):)|
(?(?true)|(?false)|(?null)|""(?(?:\\.))""|(?-?(?:0|[1-9)(?:\.0-9]+)?(?:[eE?[0-9]+)?))|
(?:)|
(?\[)|
(?,)|
(?\])|
(?{)|
(?})|
(?+)|
(?[\r\n]+)|
(?.+)
Named Capture Groups
| Group Name | Matches |
|------------------|-------------------------------------------------------------------------|
| Comment | Single‑line //… or multi‑line /…/ comments (skipped). |
| key | A JSON property key (double‑quoted string) followed by a colon (lookahead). |
| value | A JSON value – one of: true, false, null, a double‑quoted string, or a number (integer, float, or scientific). |
| bool | Sub‑group inside value for true/false (for direct parsing). |
| null | Sub‑group for null. |
| string | Sub‑group for the content inside double‑quotes (without the quotes). |
| number | Sub‑group for numeric literals. |
| value_sep | A colon : separating key and value. |
| array_open | Left bracket [. |
| array_sep | Comma , between array elements. |
| array_close | Right bracket ]. |
| object_open | Left brace {. |
| object_close | Right brace }. |
| whitespace | Horizontal whitespace (spaces, tabs) – not newlines. |
| newline | Line‑break characters (CR, LF, CRLF). |
| undefined | Any other character (should not occur in valid JSON; used as fallback). |
Important Notes
No recursion – the regex only tokenises; the parser handles nesting and depth limits.
Escaped characters inside strings (\n, \t, \", etc.) are not unescaped by the regex – the parser calls UnEscape() when _allowEscapeChars is true.
Whitespace and newlines are ignored by the parser (skipped during token iteration).
The undefined group uses .+ (not .*) to avoid matching empty positions – this prevents false positives when the scanner reaches the end of the string.
Purpose
This regex is a solid foundation for building a custom JSON lexer, parser, or tokeniser for .NET projects. Its clear separation of structural elements, comments, and whitespace makes it easy to implement lightweight, hand‑crafted parsers that do not rely on heavy external libraries. It is particularly well‑suited for small to medium‑sized files, configuration blocks, or embedded data fragments, where performance and memory footprint matter. You can adapt the token stream to your own data model, add validation, or transform the JSON on the fly – all while keeping full control over the parsing logic.